Skip to main content

Quick Start Guide

Create an Object Storage bucket, then upload data to it from the portal or an S3 CLI.

Prerequisites​

  • TIR project: access to a TIR project where your role can create Object Storage.
  • (Optional) MinIO Client (mc) or s3cmd: for uploads from your own machine. See the MinIO CLI guide or the s3cmd setup guide.

Create an Object Storage Bucket​

  1. Open the service. In the TIR sidebar, go to Storage > Object Storage.

  2. Start the form. Click CREATE BUCKET. On an empty project, use the create link on the landing page instead.

  3. Fill in the fields:

    FieldNotes
    Bucket NameA unique name for the bucket, using lowercase letters, numbers, and hyphens.
    AliasThe bucket's name inside TIR. Instances and Virtual Machines mount the bucket at /datasets/<alias>. It mirrors the bucket name on this page.
    Enable EncryptionEncrypts the data stored in the bucket. Available in the Delhi region only, and cannot be changed after the bucket is created.
    Encryption TypeShown once encryption is on.
    E2E Managed: E2E creates and holds the key, and encrypts and decrypts for you. Uploads, downloads, and mounts work as usual.
    User Managed: you keep the key and pass it with every upload, download. A lost key can't be recovered and bucket can't be mounted on TIR workloads.
    Enable Object VersioningKeeps previous versions instead of overwriting them. Can also be changed later.
    Enable Object LockApplies WORM retention to uploaded objects. Create-time only — it cannot be turned on afterwards, and it forces versioning on permanently.

  4. Create. Click CREATE. The bucket appears in the listing and moves from Pending to Ready once EOS has provisioned it.

Object lock is permanent

A bucket created without object lock cannot have it added later. To get WORM retention on existing data, create a new bucket with object lock enabled and copy the objects across.

Upload Your First Objects​

From the portal​

  1. Select the bucket in the listing.
  2. Open the Data Objects tab and click Upload Data.
  3. Set the Destination Bucket / Folder Path, choose or drag your files, and confirm.

From the MinIO Client​

The Setup tab prints ready-to-paste commands with this bucket's endpoint and keys.

mc alias set <mc-alias> <endpoint> <access-key> <secret-key>
mc cp -r ./local-data/ <mc-alias>/<bucket-name>/
The mc alias is not the TIR alias

<mc-alias> is a nickname that mc saves on your machine for this endpoint and key pair, and you can call it anything. It is separate from the bucket's Alias in TIR. The Setup tab pre-fills it with the TIR alias for convenience; choosing a different name changes nothing in TIR.

From s3cmd​

s3cmd put -r ./local-data/ s3://<bucket-name>/
s3cmd ls s3://<bucket-name>/

Mount It on a Workload​

When you launch an Instance (container), Virtual Machine, or Slurm cluster, select the bucket under the storage options. It mounts read-only:

  • On an Instance or Virtual Machine, at /datasets/<alias>.
  • On a Slurm cluster, at the mount path you enter next to the bucket, such as /data.

Next Steps​

  • Features — versioning, object lock, lifecycle rules, and replication.
  • Pricing — slab rates and worked examples.

API Reference​

Object Storage API Reference

Create and manage buckets, objects, versioning, locks, lifecycle rules, presigned URLs, and replication.

Last updated on October 19, 2026.